Cecropia schreberiana in the Luquillo Mountains of Puerto Rico.
Cecropia schreberiana Miq. is one of the characteristic tree species of the Luquillo Mountains of Puerto Rico, due to its abundance and distinctive form and to its manifest link with the hurricane-driven dynamics of Luquillo forests. Consequently, C. schreberiana has been part of some 90 ecological studies. Susan Silander, however, conducted the only study focused solely on C. schreberiana. Her own research and survey of relevant work through 1979 provide the basic account of C. schreberiana biology, emphasizing its population dynamics as an early successional species (Silander, 1979; Silander & Lugo, 1990; see also Brown et al., 1983; McCormick, 1995). Glogiewicz (in press) includes more recent information on the species in a paper on C. peltata L. However, most of the work since 1979 and much of the earlier work on C. schreberiana have never been synthesized. Also, we now know much more about disturbance and forest regeneration, so relevant to C. schreberiana, in the Luquillo Mountains (Walker et al., 1996a). Thus a current review can provide a more detailed and a more contextual description of this important species. Most interesting is how the details all cohere to create a picture of a pioneer tree whose landscape-level population dynamics reflect landscape-level dynamics of forest disturbance and recovery. The population of C. schreberiana trees is literally rejuvenated by hurricane disturbance, then matures and senesces until disturbance restarts the cycle (McCormick, 1995).
This paper has three main sections. In the first section, I review studies on the relatively fixed features of C. schreberiana biology in Luquillo forests, such as tree physiognomy, seed germination, and determinants of growth. In the second section, I use an understanding of the species' biology and of the disturbance regime in the Luquillo Mountains to interpret the dynamic features of C. schreberiana populations. I focus on how C. schreberiana responds to disturbances of different types, in order to explain its high abundance in Luquillo forests and to explain why it has mostly an even-aged population that cycles rapidly around the seed-to-tree-to-seed life history. In the third section, I discuss the species' key role in nutrient storage, and possible role as a facilitator, during succession after disturbance.
A. THE LUQUILLO MOUNTAINS
The Luquillo Mountains lie in the northeast comer of Puerto Rico and cover an area of about 27,000 ha, 11,330 ha of which are included in the Luquillo Experimental Forest (congruent with the Caribbean National Forest). The mountains rise to 1000 m within 5 km of the ocean, with five peaks topping that elevation and the highest reaching 1075 m. Prevailing winds coming off the ocean from the east drop rain as they rise over the mountains; thus rainfall increases with elevation, ranging from an average of 3537 mm [year.sup.-1] at low elevations to 4849 mm [year.sup.-1] at highest elevations (Garcia-Martino et al., 1996). February through April are the drier months but monthly rainfall is variable. Mean monthly temperatures at the highest elevations range from ca. 17 [degrees] C in January to 20 [degrees] C in September, and at lowest elevations from ca. 23.5 [degrees] C to 27 [degrees] C in those same months (Brown et al., 1983).
The Luquillo Mountains are composed mainly of igneous rock formed in the Cretaceous, with some intrusive materials from the Tertiary (Snyder et al., 1988; Weaver, 1995). The soils are classified as inceptisols and ultisols and are mostly leached acidic clays and silty clay loams. Soil drainage tends to decrease with elevation. Due to rapid decomposition, little humus accumulates, except in local areas at upper elevations, where decomposition seems to be inhibited by waterlogging.
There are four main life zones (Holdridge System) in the Luquillo Mountains: subtropical wet and subtropical rain forests are found at low and middle elevations, lower montane rain and lower montane wet forests at high elevations (Ewel & Whitmore, 1973, in Brown et al., 1983). There is also an area of subtropical moist forest at low elevations on the southwest slope.
Locally, the forest vegetation is classified into four types. The "tabonuco forest," so named for the dominant tabonuco tree, Dacryodes excelsa Vahl, covers lower slopes to about 600 m. In well-developed stands the larger trees exceed 30 m in height, there is a fairly continuous canopy at 20 m, and the shaded understory of variable structure is moderately dense. There are about 168 tree species in this forest type. The "colorado forest," named for the large colorado tree, Cyrilla racemiflora L., begins above the tabonuco forest and extends up to about 900 m. Its canopy reaches only about 15 m. There are some 53 tree species in this forest type. Soils are saturated and root mats above the soil are common. At this same elevation, but in especially steep and wet areas, is "palm forest," heavily dominated by the sierra palm, Prestoea acuminata (Willd.) H. E. Moore (= P montana: Henderson et al., 1995). Patches of palm forest are also found in saturated riparian areas at lower elevations within the tabonuco forest. The palm forest reaches about 15 m in height. At the highest elevations is "dwarf forest," a dense forest as short as 3 m, on saturated soils. Here the trees and ground are covered with mosses and epiphytes. Ascending the Luquillo Mountains through these forest types, the average tree height and dbh, number of tree species, and basal area [ha.sup.-1] tend to decrease, while stem density increases (White, 1963; Weaver & Murphy, 1990). (For fuller descriptions of forests in the Luquillo Mountains, see Wadsworth, 1951; Odum & Pigeon, 1970; Brown et al., 1983; Snyder et al., 1988; Lugo & Scatena, 1995; Lugo et al., 1995; Weaver, 1995).
Above 670 m the forest has not been disturbed by humans except locally for roads and a few buildings. Between 670 and 600 m, in the lower part of the colorado forest on the western slopes, there has been a minor amount of farming and some tree thinning for "timber stand improvement" 50 years ago. In the tabonuco forest zone, i.e., below about 600 m, there has been more human use, including farming, logging, charcoal production, and coffee cultivation. Most of this zone within the Experimental Forest has been forested since the 1930s, with little human interference. But outside the boundaries of the Experimental Forest, on the lowest slopes of the mountains, there are today areas of pasture and early second growth. (For forest history, see F. Wadsworth, 1950, 1970; Snyder et al., 1988; Scatena, 1989; Garcia-Montiel & Scatena, 1994). Natural disturbances include background treefalls (not caused by hurricanes), landslides, and hurricanes. I will describe the natural disturbance regime in a later section.
Many studies cited in this paper took place in tabonuco forest at El Verde Field Station, described in Odum and Pigeon (1970) and Reagan and Waide (1996). El Verde is in the subtropical wet forest zone; it is at about 350 m elevation, rainfall averages 3450 mm per year, and soils are upland ultisols and oxisols, mainly as well-drained clays and silty clay loams (Edmisten, 1970).
III. Biology of Cecropia schreberiana
A. THE GENUS CECROPIA
Cecropia Loefl. (Cecropiaceae) is the archetypal genus of neotropical pioneer trees, found throughout tropical America and represented by about 75 species (Mabberly, 1987). The common African pioneer Musanga cecropioides R. Br. (Cecropiaceae) is morphologically and ecologically similar to Cecropia (McKey, 1988), and the numerous Asian pioneers, Macaranga spp. (Euphorbiaceae), resemble it closely (pictures in Whitmore, 1990). As a genus, Cecropia's pioneer characteristics include abundant seed production, wide seed dispersal, seed dormancy, disturbance-cued germination, shade intolerance, fast height-growth, early maturity, and short life (Holthuijzen & Boerboom, 1982; Vasquez-Yanes & Smith, 1982; Brokaw, 1986; Alvarez-Buylla & Martinez-Ramos, 1992). Cecropia schreberiana shares these generic attributes.
B. DESCRIPTION OF CECROPIA SCHREBERIANA MIQ.
Cecropia schreberiana Miq. was lately distinguished from C. peltata L. (Howard, 1988), and thus only recent publications use the new name. Whereas Cecropia peltata occurs in Mexico and Central America, C. schreberiana occurs in the Antilles and northern South America (Howard, 1988; ISTF, 1997) and presumably is the only Cecropia species in the Luquillo Mountains. The species is known in Puerto Rico by several common names: grayumo hembra, llagrumo hembra, yagrumbo hembra, trumpet tree, trumpet wood ("trumpet" due to hollow new branches: Silander, 1979). The description of C. schreberiana given here is simplified from detailed treatments (Little & Wadsworth, 1964; Liogier, 1985; Howard, 1988).
Cecropia schreberiana in the Luquillo Mountains typically reaches 20 m in height and 60 cm dbh, but trees 25 m tall and 75 cm dbh have been recorded (F. Scatena, pers. comm.; F. Wadsworth, pers. comm.). Its size decreases as elevation increases (Weaver, 1986). Leaves of mature trees are simple, alternate (but clustered), 30-75 cm wide, with 7-11 large lobes on a long, thick petiole (Fig. 1). The leaves are silvery below and are therefore conspicuous from a distance on windy days. Seedling leaves are unlobed, lanceolate, slightly toothed, downy on both surfaces, and whitish below. Branches are few and stout and support a thin, spreading canopy. The bark is smooth and gray colored; triangular leaf scars are evident on younger branches. The tree's wood is soft, weak, and lightweight; it has the lowest specific gravity, 0.29 (Little & Wadsworth, 1964), among dicot trees in the colorado forest (Weaver, 1995), and in tabonuco forest only the uncommon balsa, Ochroma pyramidale (Cav.) Urban, is lighter (Odum, 1970b). Stilt roots descend from about 1 m up the trunk. These are superficially rooted in the ground (Silander & Lugo, 1990). The roots have endotrophic mycorrhizae (Lodge, 1996; Calderon, 1993, in Myster & Walker, 1997; not ectotrophic mycorrhizae, as reported by Edmisten ).
[Figure 1 ILLUSTRATION OMITTED]
The "most characteristic" size and form of a "normally full-grown or mature individual" of C. schreberiana, in a tabonuco forest about 35 years after hurricane damage, was a tree 15 m tall and 23.6 cm dbh, with a trunk height to first branch of 7 m and a crown 12 m wide (based on 53 trees [is greater than or equal to] 10 cm dbh in eight 10 x 100 m quadrats [Holdridge, 1970]). But this "most characteristic" form was the image in just one frame of a movie; as we shall see, size-class distributions of C. schreberiana change greatly with time after disturbance.
Cecropia schreberiana is dioecious. Flowers of both sexes are grouped on clustered spikes (aments) (Fig. 1) and are minute (ca. 1.6 mm long) and numerous (mean of 15,140 per pistillate cluster) (see Silander, 1979, for more details). The female spikes develop into multiple fruits, swollen to 5-10 cm long and 1 cm thick and containing many minute fruits, each with one seed (achene). The small oblong seeds are about 2 mm in length. Mean wet mass of the seed is 0.0029 g (Everham et al., 1996), of which 68.8% is water. Estimates of mean dry weight range from 0.0003 to 0.00098 g (ITF, 1959; Silander, 1979; Devoe, 1989; Everham et al., 1996; Myster, 1997). Cecropia schreberiana has the smallest seeds among canopy trees in the colorado forest (Weaver, 1995) and had the next to smallest, after Homalium racemosum Jacq., among 29 canopy tree species in a tabonuco forest (Smith, 1970b). In an extraregional comparison, its seeds would be at the lighter end of the range in seed weights of North American weeds (Silander, 1979).
Unlike its mainland congeners, C. schreberiana in most of the Antilles does not have symbiotic ants that protect it from herbivores and burdensome vines, and it lacks the trichilia (glandular areas) at the base of petioles that its congeners have for producing Mullerian bodies to reward ant defenders (Janzen, 1973; Rickson, 1977).
Various other information on physical characteristics of C. schreberiana is available: pollen (Ogle, 1970); seedlings (Duke, 1970a); leaf form (Smith, 1970a); leaf biomass (203 g [m.sup.-2] leaf), leaf weight/area ratio (8.01 mg [cm.sup.-2]) and leaf ash content (Odum, 1970b); specific leaf area (a low 43.8 [cm.sup.2] [gm.sup.-1]: La Caro & Rural, 1985); leaf sclerophylly, fiber, and lignin (all are high: La Caro & Rudd, 1985); chlorophyll A in leaves (Odum & Cintron, 1970); wood and vascular system (Odum et al., 1970b); biomass of various plant parts in different life stages (Silander, 1979); biomass and nutrient content of leaves, branches, wood, bark, and roots (Ovington & Olson, 1970; La Caro & Rudd, 1985; Bloomfield, 1993 in Schowalter, 1994; Scatena et al., 1993, 1996); and relationships between trunk diameter and bark thickness, and between trunk and crown diameters (Murphy, 1970). Glogiewicz (in press) summarizes information on propogating C. schreberiana and its woodworking qualities and uses (see also Little & Wadsworth, 1964; Silander & Lugo, 1990).
C. DISTRIBUTION IN THE LUQUILLO MOUNTAINS
In the Luquillo Mountains Cecropia schreberiana is found as juvenile trees in recently disturbed areas, as trees of all ages on permanent edges such as roadsides and stream banks, and as mature trees in older forest (Crow & Grigal, 1979; Silander, 1979). It occurs in nearly all forest types in the Luquillo Mountains (Weaver, 1994), being most abundant at mid-elevations in the tabonuco forest (e.g., Briscoe & Wadsworth, 1970), moderately abundant in colorado and palm forests (Weaver, 1983, 1986), absent in the dwarf forest except along roads and as a rare colonizer in other human disturbances there (Byer & Weaver, 1977; Weaver, 1990), and surprisingly uncommon in abandoned pastures at all elevations (Aide et al., 1995, 1996; Zimmerman et al., 1995).
In colorado forest Cecropia schreberiana occurs mostly in valley bottoms, possibly because windthrow creates gaps suitable for colonization more often in valleys than on slopes and ridges (Weaver, 1986). But in tabonuco forest there was no relationship between C. schreberiana importance value and steepness of slope (R. Wadsworth, 1970). In terms of exposure, it is equally abundant in colorado forest on windward and leeward slopes of El Yunque Peak (Weaver, 1991, 1995).
There is much inventory data on the abundance and size-class distribution of Cecropia schreberiana in particular forest stands at particular times. The species can be among the ten most abundant species in relatively mature tabonuco forests (Wadsworth, 1951; Briscoe & Wadsworth, 1970; Crow & Grigal, 1979). For example, in about four hectares each of "virgin" tabonuco and colorado forests, Wadsworth (1951) recorded 83 and 17 C schreberiana per ha with dbh = 4 in. (~10 cm), and he mentioned it as common in palm forests. Additional data on abundance and size-class distribution of C. schreberiana in Luquillo forests are available for tabonuco forest (R. Wadsworth, 1970; Smith, 1970b; Soriano-Ressy et al., 1970; Silander, 1979; Crow, 1980; Weaver, 1983, 1994; Dallmeier et al., 1992; Zimmerman et al., 1994; Scatena & Lugo, 1995; Zou et al., 1995), colorado forest (Weaver, 1983, 1986, 1989, 1991, 1994, 1995), palm forest (Weaver, 1983; Lugo et al., 1995), and dwarf forest (Weaver et al., 1986; Weaver, 1994). Cecropia schreberiana and other poor timber species were removed from some Luquillo forest stands in 1938 and 1939 to improve naturally established stands (F. Wadsworth, 1970). Little result was noted, since, it was reasoned, these species had been only a minor stand component in numbers and effect on other species.
As a pioneer tree, C. schreberiana's population abundance and size-class distribution at any given place and moment depend strongly on the time elapsed since major disturbance, and so to interpret the static data of these inventories one must first understand this species' life history.
D. FLOWERING, FRUITING, AND SEED PRODUCTION
Cecropia schreberiana flowers as early as 3.3 years of age in permanently open, sunny arcas such as roadsides; flowering occurs later, but as early as 5.6 years, in forest gaps where light is less and trees would put more energy into height-growth to compete for it (Silander, 1979; Silander & Lugo, 1990). Over a reproductive lifetime a female tree could produce nearly 38,000,000 flowers (using 30 years for longevity: Silander, 1979). Male trees flower earlier than females, perhaps because pollen is comparatively cheap to produce. Cecropia schreberiana is apparently wind-pollinated (McCormick, 1995).
Flowers and fruits of C. schreberiana can be seen throughout the year but vary seasonally in abundance. Estrada Pinto (1970) recorded flowers mainly from October through March. Silander (1979) observed a flower and fruit peak from January to March, the annual period of minimum temperatures and rainfall and shortest days, and she recorded the fewest flowers and fruits from June to September. Murphy's (1970) observations of 25 mature trees for about two years revealed no temporal pattern. Flowering and fruiting is more seasonally restricted in younger trees, probably because they must reserve energy for height-growth (Silander, 1979). The male inflorescences remain on the tree for about 1.5 months, producing abundant pollen for 1-1.5 months. Fruits mature in 3.5-4 months, and the next flowering begins about 1.5 months later (Silander, 1979).
The flowers produce about 18% viable seed. This mounted to a mean of 474,185 and maximum of 1,090,080 viable seeds per tree among 50 mature trees during one year of observation (Silander, 1979). Seed production increases during the life of a tree (until senescence begins), and, based on estimates of age-specific reproductive output, a tree can produce 6-7 million viable seeds over a lifetime (Silander, 1979; Silander & Lugo, 1990). Combining age-specific reproductive output and the size-class distribution of C. schreberiana at one point in time (probably several decades after a hurricane) Silander (1979) estimated that the C. schreberiana population at that time provided 731 seeds [m.sup.-2] [year.sup.-1], with 375 [m.sub.-2] [year.sup.-1] available for germination. Of course, these figures would vary greatly depending on the changing status of the C. schreberiana population after a disturbance.
Numerous bird species and at least two bats are the primary dispersers of C. schreberiana seeds in Luquillo forests. Leck (1972) noted 20 bird species, including six migrant species, visiting C. schreberiana and saw 14 of them feed on its fruits. These species included almost all the common birds of the Luquillo Mountains. Cecropia schreberiana fruit is the main food of the two bats Artibeus jamaicensis and Stenoderma rufum (Willig & Gannon, 1996). From knowledge of A. jamaicensis's metabolic needs and the fact that C. schreberiana fruit contains 4,675 calories [g.sup.-1] dry weight (Scogin, 1982 in Willig & Gannon, 1996), it was calculated that an adult A. jamaicensis must eat at least 11.2 g dry weight of C. schreberiana fruit each night to maintain a positive energy balance (Willig & Gannon, 1996). The introduced black rat (Rattus rattus) eats C. schreberiana fruits, which it prefers about equally to other fruits (Weinbren et al., 1970); C. schreberiana seeds were found in stomachs of the lizard Anolis evermanni (Reagan, 1996), while Silander (1979) once saw an Anolis eating the fruit; and some seeds of the tree are undoubtedly dispersed by gravity (Silander, 1979) and wind.
Studies on seeds of Cecropia peltata L. have shown that passage through bird digestive tracts may enhance germination (Olson & Blum, 1968) and that passage through bats definitely does (Fleming & Heithaus, 1981). Cecropia fruits may be dispersed by bats up to several kilometers from a source tree (Fleming & Heithaus, 1981). When defecated by bats Cecropia seeds can be spread in a trail up to 4 x 0.3 m (Charles-Dominique, 1986), but clumped seeds and seedlings often result (Fleming & Heithaus, 1981).
Seed dispersal of Cecropia schreberiana is widespread relative to the density of adult trees and to the dispersal ability of other tree species in tabonuco forest (J. Zimmerman, unpubl. data). Concerning dispersal of this species to open versus closed vegetation, results from different studies appear to contradict each other At El Verde 36 times as many seeds fell in three human-made gaps than in nearby closed forest (16.1 versus 0.46 seeds [m.sup.-2] [year.sup.-1]: Devoe, 1989). By contrast, in small samples within and at the edge of a forest at least twice as many C. schreberiana seedlings germinated in sterilized soil flats than in similar flats in an adjacent landslide (Walker & Nens, 1993). Cecropia schreberiana was the second most common animal-dispersed tree seed falling in the human-made gaps and fourteenth most common in the closed forest (Devoe, 1989). It was one of the more common tree seedlings in the forest-landslide study (Walker & Neris, 1993).
F. SOIL SEED BANK
On the basis of germination of seedlings from soil samples, Bell (1970) concluded that the most abundant seeds in the soil in the tabonuco forest at El Verde were Cecropia schreberiana. Its seedlings made up 60.5% of seedlings germinated from the soil samples, even though adults of the species only made up 4.6% of the trees in the canopy of the studied forest (Smith, 1970b). Using two methods, Silander (1979) estimated there were 342 and 409 C. schreberiana seeds [m.sup.-2], 3-5 cm deep in the soil at El Verde. In another study both landslide and forest soil samples were dominated by seeds of this species (Guariguata, 1990). Again, we must remind ourselves that these densities will change in step with changes in the C. schreberiana population.
High densities of Cecropia schreberiana seeds accumulate in the soil because its seeds are abundant and capable of dormancy. Seeds in moist peat in a laboratory remained viable for at least six months, and those on the forest floor remained viable for two to three months (Silander, 1979). In another experiment, the viability of C. schreberiana seeds collected and stored at 5 [degrees] C was 60% in March (when collected) and 53% the following October (Walker, 1994). Viability declines, of course, particularly between the time when fruit is ripe on the tree and after it reaches the ground; 71% of seed from freshly collected fruit germinated, whereas only 18% from fallen fruit did (Silander, 1979). The C. schreberiana seed bank is also reduced by insect attack, especially by Nitidulidae (Silander, 1979) and probably ants (Myster, 1994), and by pathogens. In a study of insect and pathogen attacks on C. schrcberiana seeds collected from trees and then placed in a landslide, Myster (1997) found that 10% were predated and 2% were affected by pathogens before dispersal, while an additional 15% were predated and 15% lost to pathogens after dispersal. Pathogen attack was greater at the edge than in the center of the landslide (Myster, 1997). Thus, while many seeds do remain viable for several months, high densities of germinable seeds in the soil must be sustained by repeated additions (Silander & Lugo, 1990).
Research over the past three decades has shown with increasing precision that Cecropia schreberiana seeds germinate when the forest canopy is removed by disturbance, and that germination is positively associated with light and temperature and negatively associated with litter and saturated soil (Bell, 1970; Duke, 1970b; Silander, 1979; Devoe, 1989; Guzman-Grajales & Walker, 1991; Everham et al., 1996). The congener C. peltata germinates in response to the high ratio of far-red to red light found in gaps (Vasquez-Yanes & Smith, 1982). Germination of C. schreberiana is epigeal (Silander & Lugo, 1990).
An experiment at El Verde included a site from which vegetation had been removed from 320 [m.sup.2], a site irradiated with cesium that removed or affected canopy from 5000 [m.sup.2], and a closed-canopy control with no treatment (Odum, 1970a; Brown et al., 1983). There were no Cecropia schreberiana seedlings present in these sites beforehand (Duke, 1970b; McCormick, 1970). Within a few months after treatment there was a flush of newly germinated seedlings of C. schreberiana in the center of the vegetation removal site (1.74 seedlings [m.sup.-2], 19.4% of emerging seedlings), fewer at the irradiated site (0.12 seedlings [m.sup.-2], 5.2% of seedlings), and none in the control (Duke, 1970b). There were fewer seedlings in the irradiated site perhaps because the canopy opened gradually and seeds may have been killed by irradiation.
Transfer experiments confirmed these results. Two equal-sized samples of soil were moved from under forest canopy: the first to the 320 [m.sup.2] open area mentioned above and the second to another site beneath the canopy. In the first, 271 Cecropia schreberiana seedlings germinated (181 seedlings [m.sup.-2] of soil); in the second, 8 seedlings germinated (Bell, 1970). Seedlings had begun to germinate after two weeks. Silander (1979) elaborated on these experiments. She moved forest soil samples to a forest gap, an open field, and to another forest site, and shielded them from further seed input; she found three times as many C. schreberiana seedlings in the gap (87 seedlings [m.sup.2]) than in the open field, and no seedlings under forest canopy. Germination in this experiment peaked at four weeks. In a later study C. schreberiana germination was positively associated with light and temperature (Everham et al., 1996), both of which are relatively high in gaps (Silander, 1979). Under favorable conditions germination of C. schreberiana can reach 80-90% (ITF, 1959; Silander & Lugo, 1990).
In landslide situations, however, plant cover, not openness, can enhance germination of Cecropia schreberiana (Myster, 1997). Seeds sown under fern (Dicranopteris pectinata) thickets in landslides germinated at a higher rate than those sown in adjacent open areas of landslides (Walker, 1994). Under these thickets total soil N is higher than in the open areas, making thickets favorable sites for C. schreberiana, a highly nutrient-demanding species (Walker et al., 1996c). Soil moisture was higher (saturated soil, however, inhibits germination: Everham et al., 1996), and soil bulk density and soil surface temperatures were lower, in the thickets than in the open areas. Half the maximum germination was normally reached within 2-3 weeks of sowing. As may happen in open fields (Silander & Lugo, 1990), germination could have been hindered in the open landslide by extreme fluctuating temperatures. In open fields, germination is inhibited by liner and perhaps by high daytime soil temperature, fluctuating wet-dry conditions, and a hard soil surface (Silander, 1979).
Leaf litter may hinder germination and/or emergence of Cecropia schreberiana seedlings by preventing light from reaching the soil, preventing increases in soil temperature, containing a chemical inhibitor, and/or being a physical barrier to seedling penetration. Germination in a recently created gap was only 2% for seeds sown on top of litter, but increased to 30% for seeds sown where litter was removed (Silander, 1979). After Hurricane Hugo blew large quantities of leaves to the ground, similar results were seen among treatments in which litter was removed (highest germination), not changed, or added (Guzman-Grajales & Walker, 1991; and see Everham et al., 1996).
Experiments to detect possible chemical inhibition of germination produced seemingly contradictory results. Seeds placed in petri dishes (protected from litter) under forest canopy did not germinate, suggesting that liner chemicals were not the cause (Silander, 1979). But in a laboratory, 26% of seeds under aluminum foil (presumed no chemical effect) germinated and 0% under leaf litter germinated, suggesting that the litter had a chemical effect (Silander, 1979).
H. GROWTH AND METABOLISM
Cecropia schreberiana is a light-demanding tree that grows rapidly until it reaches the canopy in Luquillo forests. All 108 trees in transect studies in tabonuco forest were receiving full top light as canopy trees in the forest, in gaps, or on roadsides (Silander, 1979; and see Murphy, 1970). Dominant individuals grow much faster than codominant, intermediate, and suppressed trees (Crow & Weaver, 1977). (Wadsworth  presented data suggesting that intermediate and suppressed trees grow fastest, but the small sample size and variable initial tree size confound interpretation of the data [cf. Wadsworth, 1957].) When exposed to adequate light and nutrients, C. schreberiana grows rapidly and can reach 20 cm dbh in four years from germination (Scatena et al., 1996). It has high values for maximum rate of net photosynthesis, light saturation point, compensation point, and ratio of net photosynthesis to night respiration (Lugo, 1970; Odum et al., 1970a). Further information on C. schreberiana growth is organized below by seedling (arbitrarily, to ca. 1 m tall), sapling/pole (1-10 m), and mature tree ([is greater than or equal to] 10 m) stages.
Seedlings of Cecropia schreberiana are especially light demanding (Silander & Lugo, 1990). In a transect study no C. schreberiana seedlings were found beneath the forest canopy (Silander, 1979; and see Devoe, 1989). In the open, height-growth rapidly transforms seedlings into saplings, as measured by different authors: an average of 2.1 m in the first year (Marrero, 1954 in Silander, 1979); 10.16-15.24 cm in 10 weeks in prepared seedling beds, and to 1.83 m in 7 months when transplanted to favorable sites (ITF, 1959); and to 0.76 m (mean 0.49) in the first 8 months, permitting a maximum height of 1.14 m (mean 0.73) in a year (conditions not specified) (Silander, 1979). The height-growth pattern of seedlings in their first 8 months (July to February) was a sigmoidal-shaped curve (Silander, 1979), suggesting an inherent growth pattern or perhaps a response to seasonal climate. Growth was highest in September, when days are long and warm, then the curve leveled off during the drier and cooler part of the year.
Other measures of growth are stem diameter, which can reach 0.5 cm (mean 0.36) in the first 8 months (Silander, 1979), and number of leaves. Duke (1970b), upon finding more leaves on Cecropia schreberiana seedlings in the cut than in the irradiated site, suggested that seedlings in the cut were older and/or in a better microenvironment (recall that canopy opening was gradual in the irradiated site).
Cecropia schreberiana seedlings achieve fast growth with a maximum photosynthetic rate of 0.30 g C [m.sup.-2] [hr.sup.-1], an average saturation point of 0.30 g-cal [cm.sup.-2] [minute.sup.-1], an average compensation point of 0.020 g-cal [cm.sup.-2] [minute.sup.-l], and a net photosynthesis-to-night respiration ratio of 4.7 (Lugo, 1970). In contrast, the same parameters for the dominant tabonuco tree (Dacryodes excelsa) were 0.06, 0.05, 0.007, and 0.79, respectively. Leaves of "very young trees" (presumably seedlings and saplings) respire faster than leaves of older trees (Odum et al., 1970a). (Lugo [ 1970] gives additional details on metabolism.)
Two studies already described showed that Cecropia schreberiana seedlings do not necessarily grow best where germination is best. In the first study, seedlings emerged faster where leaf litter was cleared than where it was unchanged or added, but mean height-growth was faster in the litter-unchanged and litter-added treatments (Guzman-Grajales & Walker, 1991). Light levels were presumably similar across treatments, so the growth differences may have been due to higher nutrient levels in the litter-unchanged and litter-added treatments (fertilization does enhance establishment of C. schreberiana: Walker et al., 1996c). In the second study, germination was higher under fern thickets in landslides, but, once established, older seedlings tended to grow faster (nonsignificantly) in open areas of landslides (Walker, 1994), perhaps in response to higher light there.
An experiment using Cecropia schreberiana seedlings transplanted into a landslide and treated with N and P confirmed that this species' seedlings respond dramatically to added nutrients (Fetcher et al., 1996). Biomass increased significantly with added N and P, and addition of N also increased light saturated photosynthesis in open sites, helping explain why biomass increased more in open sites than in edge sites when both N and P were added. Addition of N and P tended to increase foliar concentrations of those nutrients and had variable results on allocation of N and P to roots versus leaves. Based on its performance without added nutrients, Fetcher et al. (1996) concluded that C. schreberiana was not well adapted to colonizing exposed parent material.
As with seedlings, C. schreberiana saplings and poles also are not found beneath closed canopy (R. Wadsworth, 1970). But in the open, rapid height-growth continues in the sapling stage, to 2.16 m (mean 0.79) [year.sup.-1] in height and to 3 cm (mean 0.65) [year.sup.-1] in diameter (most material in this section from Silander, 1979). However, Odum et al. (1970b) mention height-growth from 1 to 7 m in height in 17 months, or 4.23 m [year.sup.-l], in the vegetation removal site mentioned above. New leaves are produced about every two weeks and stay on the tree for 3.5-4 months. Thus leaf sears allow aging in this stage. The ratio of photosynthetic to non-photosynthetic tissue decreases and levels off at about 2 cm diameter and 2 m height, i.e., at about 2.5 years. Height-growth dominates diameter-growth in the first 4-5 years (Silander & Lugo, 1979), but form depends partly on environment: trees in full light on roadsides branch at about 3.3 years; those in forest gaps, with mainly top light, branch at about 5.6 years.
3. Mature Trees
After about 10 years of fast growth in seedling, sapling, and pole stages, growth of Cecropia schreberiana slows when trees reach mature size and a position in the canopy (Silander, 1979). Silander and Lugo (1990) reported a mean annual diameter increment of 0.64 cm among 50 "mature dominant trees" and a maximum of 1.52 cm, while Crow and Weaver (1977) recorded 0.51 cm (range 0.08-1.55 cm) diameter increment per annum among 49 trees with a mean dbh of 31.7 cm. But diameter-growth declines among trees [is greater than] 9 cm dbh (Crow & Weaver, 1977), and basal area increment, as a function of the area of tree crown, peaked in one study at a crown area of about 80 [m.sup.2], then dropped at higher values (Murphy, 1970). As a mature tree, C. schreberiana is, in fact, one of the slower growing trees in tabonuco forest (Murphy, 1970; Crow & Weaver, 1977). Murphy (1970) recorded a mean annual growth of 0.2 cm among 50 mature C. schreberiana trees (mean dbh 22 cm), as compared to 0.36 and 0.32 em for equal numbers of mature Dacryodes excelsa and Manilkara bidentata (A. DC.) Cher., respectively. The slow growth of mature C. schreberiana may reflect the requirements of copious fruit production and/or the limited photosynthetic area of this species' thin canopy (Murphy, 1970). In fact, leaves represent a reduced percentage of total tree biomass in larger size classes (Silander, 1979). However, the relatively few leaves on a mature tree are still photosynthesizing rapidly. Asymptotic leveling of net photosynthesis in crown leaves is higher in C. schreberiana than in Dacryodes excelsa, and ratios of total net photosynthesis of the leaves in daytime to their night respiration were 12 in C. schreberiana, 10 in Prestoea acuminate, and 9.4 in Dacryodes excelsa (Odum et al., 1970a).
Some other generalizations and numerous miscellaneous results concerning growth and metabolism of Cecropia schreberiana can be found in the literature: There is no correlation between tree rings and annual growth (Odum et al., 1970a). Growth rates of pioneer trees (including C. schreberiana) decline with increasing elevation (Weaver, 1986). A correlation of C. schreberiana growth rates during defined periods with rainfall before and during those periods showed that the rate peaked at one-fourth maximum rainfall (Murphy, 1970). Cecropia schreberiana grew nearly three times as fast in regenerating forest on a recently cutover area than in an undisturbed stand (Wadsworth, 1957, 1958). Presumably, the C. schreberiana benefited from reduced density in the cutover stand, but in another study of variably treated plots at several locations, growth of this species in thinned plots was only 14% greater than in control plots (Weaver, 1983). Cecropia schreberiana leaves fell at a rate of 27.8 [+ or -] 8.2 g [m.sup.-l] [yr.sup.-1] in ~70-year-old forest and at 18.2 [+ or -] 13.3 g [m.sup.-1] [yr.sup.-1] in mature (apparently at least three times older) forest in the tabonuco zone (Zou et al., 1995). During one year of records, leaf fall peaked in March through May and again in November (Zou et al., 1995). (For additional information on C. schreberiana growth, see Murphy, 1970; Crow & Weaver, 1977; Weaver, 1979, 1983. For more information on photosynthesis and respiration, see Odum, 1970b; Odum et al., 1970a. For information on C. schreberiana growth at Toro Negro, Puerto Rico, see Weaver, 1979.)
I. MORTALITY AND LONGEVITY
Cecropia schreberiana suffers substantial background (not due to disturbance) and catastrophic (due to major disturbance) mortality (cf. Lugo & Scatena, 1996) in Luquillo forests. The flush of germinating C. schreberiana seeds after canopy opening is quickly followed by massive seedling death and the survival of only a tiny fraction that will reach maturity. In the definitive study of C. schreberiana survivorship, Silander (1979) plotted survivorship as a function of age and life cycle stage (Fig. 2). The survivorship curve is steeply concave. Within a year, 99.7% of the seedlings die, even in open areas. The rate of mortality declines slightly at 1 year (early sapling stage), starts tending toward level at 3-4 years (middle sapling), and becomes near level at about 14 years (canopy).
[Figure 2 ILLUSTRATION OMITTED]
Seedling survivorship is better in gaps than under forest canopy and is poor in open fields (Silander, 1979); it tended to be lower in litter-added than in litter-removed and litter-unchanged treatments (Guzman-Grajales & Walker, 1991); and it is associated positively with light and negatively with soil moisture (Everham et al., 1996). Overtopped seedlings are the most likely to die (Silander & Lugo, 1990); however, in landslides, survivorship is better under fern canopies than in the open during the earliest stages of establishment, then better among plants in the open several months later (Walker, 1994). This reversal parallels the contrast between establishment and growth in the same environments. Presumably, landslides are such harsh environments that even this very light-demanding pioneer has difficulty colonizing the open areas, but once established does benefit from the high light there.
In the sapling stage interspecific competition is a principal cause of death, as is vine load and herbivory (Silander, 1979). Cecropia schreberiana in thickets colonizing the gaps created by Hurricane Hugo in 1989 suffered higher herbivory (11% leaf area missing) than did isolated saplings (5% leaf area missing) or larger trees of the species, and damage increased during 1991-1995 (Schowalter, 1994, 1997, pers. comm.). Among C. schreberiana colonizing landslides, 25% and 2% of leaf area was lost to herbivores and pathogens, respectively (R. Myster, unpubl, data). Diversity of herbivorous insects on its saplings is not particularly high (Schowalter, 1995).
Among larger trees termites, vines, and soil slippage are blamed for mortality, and Silander (1979) speculated that the adults compete poorly for soil water and nutrients in the closed forest. She also suggested that C. schreberiana's weak wood made canopy adults relatively susceptible to wind damage, as was borne out by studies on the effects of Hurricane Hugo. During Hugo this species suffered especially high levels of stem breakage and mortality at El Verde, and the larger, and presumably more exposed, the tree, the greater the mortality (Zimmerman et al., 1994). Of 136 C. schreberiana [is greater than or equal to] 10 cm dbh in the 16 ha Hurricane Recovery Plot at El Verde, 18.4% were uprooted, 21.3% had trunks snapped, and 16.9% of remaining trees suffered branch damage by Hugo, versus means of 9.7%, 8.1%, and 25%, respectively, for all other common species (Zimmerman et al., 1994). Mortality of C. schreberiana was 52.9% about 1 year after the hurricane, while only 31.3% had sprouted, versus 8.4% and 65.3%, respectively, for all other common species. But at another location, not in the Luquillo Mountains and where Hugo's wind velocity was lower, C. schreberiana lost its leaves, apparently reducing wind resistance, and suffered comparatively minor damage (Francis & Gillespie, 1993). Bates (1929) observed that many C. schreberiana in the Luquillo District were uprooted by the Hurricane of 1928. The species is shallowly rooted and (juveniles especially) is easily tipped up (Silander & Lugo, 1990). Irradiation by cesium killed C. schreberiana (Smith, 1970b), and the dead trees were especially prone to beetle infestation (Odum et al., 1970b; Smith, 1970b).
It is estimated that Cecropia schreberiana reaches the canopy in 10-12 years (Silander, 1979) and survives there until about age 30 (Wadsworth, pers. comm., in Odum et al., 1970a) to 50 (Doyle, 1981). Silander (1979) aged saplings using leaf sears and aged mature trees using probable times of establishment after dated disturbances. Using data on dbh and height of these trees, she calculated linear regressions to ascertain age, both from dbh (age [years] = 0.90 + 0.86 dbh [cm]) and from height (age [years] = 0.91 + 1.3 height [m]). However, C. schreberiana of different sizes can be found growing near each other in gaps (Perez Viera, 1986) and post-hurricane sites (N. Brokaw, unpubl, data), the locations suggesting that the different-sized trees are even-aged, and Silander (1979) points out that her age regressions, derived from mean values in C. schreberiana stands, may better predict stand age than individual stem age. Using growth rates from Crow and Weaver (1977) and its maximum observed size, Doyle (1981) estimated that >. schreberiana reached 50 years of age.
IV. Dynamics of Cecropia schreberiana
The details of the life cycle of C. schreberiana in the Luquillo Mountains are those of a typical pioneer tree species: frequent and abundant flowering and seed production; widespread seed dispersal; establishment of a soil seed bank; germination triggered by canopy opening; fast height-growth when in high-light and high-nutrient environments; heavy mortality, producing a steeply concave survivorship curve; and a short life span that is compensated by early reproduction and (to bring the life-history strategy full circle) the abundant reproductive capacity just mentioned (McCormick, 1995). In fact, C. schreberiana is overall the most successional species among canopy species in the tabonuco and colorado forests, based on an index composed of seed size, wood specific gravity, and the ratio of seedlings and understory stems to canopy stems (Smith, 1970b; Weaver, 1995).
Given this set of pioneer characteristics, the establishment of C. schreberiana will necessarily be patchy in space and episodic in time, since establishment depends so strongly on patchy and episodic disturbance. It also follows that the C. schreberiana population in an area with a shared history of disturbance will mainly be a single-aged cohort, originating at the time of the last suitable disturbance. Lacking further disturbance and recruitment, this population will age, senesce, and drastically decline in number, at least in tree form, for in the dormant seed form the population may remain high and ready to produce a new and numerous cohort following canopy disturbance.
Against this background of its biology and population ecology, I next review studies on the connection between C. schreberiana population dynamics and different types of disturbances in the Luquillo Mountains--human disturbances, landslides, "background" treefalls (not caused by hurricanes), and hurricanes--and I will argue that only hurricanes can account for the relatively high abundance of C. schreberiana. The species is also regularly found along streamsides, where its dynamics have not been studied, although Frangi and Lugo (1991) mention extensive regeneration along steep banks of streams in floodplain palm forest after Hurricane Hugo.
A. HUMAN DISTURBANCES
Human disturbances have mainly included road construction, clearing for pasture and crops, creation of coffee plantations, and logging (Garcia-Montiel & Scatena, 1994). Cecropia schreberiana is a common colonizer in open areas along roadsides in tabonuco forest (Silander & Lugo, 1990), and it was among the most frequent woody species along a road in lower-elevation dwarf forest (Byer & Weaver, 1977). However, it is unimportant in succession in other human disturbed areas in dwarf forest (Byer & Weaver, 1977; Weaver, 1990) and is unimportant in abandoned pastures at all elevations (Aide et al., 1995, 1996; Zimmerman et al., 1995).
In a study of 23 former pastures abandoned 9.6 to 60 years previously and located on the northeast slope of the Luquillo Mountains at 10-450 m elevation, C. schreberiana was absent or had low importance values in young and intermediate-aged sites (0-60 years), and was most abundant in older sites ([is greater than] 60 years) (Aide et al., 1996). There are several possible reasons why C. schreberiana does not readily colonize pastures, even where there are seed-producing adults nearby. As previously mentioned, germination in open fields is inhibited by litter and ground cover and perhaps by high daytime soil temperature, fluctuating wet--dry conditions, and a hard soil surface (Silander, 1979). Establishment may be hindered by dense herbaceous vegetation in pastures and fields, and survival of seedlings transplanted to open fields is low (Silander, 1979). Other potential problems include deficient nutrients in pastures (Zimmerman et al., 1995); lack of disturbed soil, which, for example, promotes C. schreberiana establishment at tip-up sites (L. Walker, unpubl, data); and, at least in lowland pastures, lack of ant protection against relatively greater lowland pressure from herbivores and vines (Janzen, 1973).
I do not know of any study of succession after annual crop raising in the Luquillo Mountains that could indicate C. schreberiana's ability to colonize previously farmed areas. However, it readily colonized abandoned coffee plantations damaged by Hurricane Hugo, reaching higher importance values in these plantations than in adjacent hurricane-damaged forests that had previously been selectively logged (Zimmerman et al., 1995). In fact, high density of C. schreberiana after Hugo was a good indicator of old coffee plantations in the Luquillo Mountains (J. Zimmerman, pers. comm.). Greater establishment in old coffee plantations may be supported by high levels of N, a relic of the N-fixing tree Inga vera Willd., planted to shade and fix N for the coffee plants but now mostly absent from the stands. As for logging, in a 1 ha sample of "mid-successional forest," developing since the 1920s after clear-cutting tabonuco forest for timber, C. schreberiana was the third most abundant species, with 32 trees [is greater than or equal to] 10 cm dbh, and second highest in basal area, with 3.3 [m.sup.2] (Zou et al., 1995). In a "recently cut over" site of 1.2 ha there were 50 C. schreberiana up to 10 cm dbh (Wadsworth, 1958).
Cecropia schreberiana's response to the irradiation treatment described above has been well documented. The relative density of its seedlings went from 0% before the cut to ca. 10% within two years (Lebron, 1977 in Brown et al., 1983). Among saplings (erect woody plants [is greater than or equal to] 1.4 m tall), C. schreberiana was the second most important species in the site for the first three years, its relative density peaking three years after treatment at 21.1%; it then declined gradually to 1.4%, 25 years after irradiation (Taylor et al., 1995). In the cut site C. schreberiana colonized quickly (Duke, 1970b; Odum et al., 1970b), but there was no further study.
Cecropia schreberiana is one of the dominant trees in regrowth on landslides (Walker et al., 1996b; Myster & Walker, 1997). Landslides are natural, but often human-induced, disturbances. They are estimated to denude between 0.08% and 1.1% (Guariguata, 1990; Larsen & Torres-Sanchez, 1992) of the forest area per century and to turn over the canopy every 1250-1300 years (Guariguata, 1990; Scatena & Lugo, 1995). The distribution of slides across the landscape is uneven and depends greatly on substrate and local topography (Guariguata, 1990; Larsen & Torres-Sanchez, 1992). However, 53% of recent slides are related to road construction (Guariguata & Larsen, 1990 in Walker et al., 1996b). Median area of slides was 175 [m.sup.2] in the Bisley watersheds and 161 [m.sup.2] in the Luquillo Experimental Forest (Larsen & Torres-Sanchez, 1992; Scatena & Lugo, 1995; for a review of the landslide disturbance regime, see Walker et al., 1996b).
As mentioned previously, Cecropia schreberiana seeds dominate soil seed banks in landslides (Guariguata, 1990). However, it effectively invades landslides only after initial colonization by nitrogen-fixing plants (Myster & Walker, 1997). It is associated in central areas of landslides with other species that are shade intolerant and facultatively mycorrhizal (Myster & Walker, 1997), but stable, high-nutrient soils in lower areas of slides are especially favorable to the species (Walker et al., 1996b). Eventually, C. schreberiana becomes one of the most common tree seedlings and saplings in landslides (Guariguata, 1990; Walker & Neris, 1993). It was by far the most abundant woody species [is greater than or equal to] 1 m tall in a study of 16 landslides ranging from 4 to 20 years old (Myster & Walker, 1997).
A dense canopy of C. schreberiana can last for 3-6 years in landslides, inhibiting colonization of other tree species (Walker et al., 1996b). In an age series of landslides the importance value of C. schreberiana among trees [is greater than or equal to] 1 cm dbh went from values in the range 3-60 at [is less than] 1 year, to 29-58 in years 16-24, to 3-7 in year 52 (for full chronosequence, see Guariguata, 1990).
C. BACKGROUND TREEFALLS
In tropical forests pioneer trees such as C. schreberiana frequently regenerate in "background" treefall gaps (Denslow, 1980)--that is, gaps caused by treefalls that are not associated with hurricanes or other large disturbances. Moreover, pioneers generally survive and grow best in larger gaps, those greater than 150-200 [m.sup.2] in area, where there is relatively more light (Brokaw, 1985). (In the present essay a gap is defined, when newly created, as a hole in the forest vegetation extending through all levels to near the ground, and gap area is taken as the area beneath the hole in the canopy.)
In the tabonuco and colorado forest, Perez Viera (1986) studied regeneration in 15 gaps ranging from 34 to 322 [m.sup.2] in area and 4 to 10 years in age. She found C. schreberiana in four of these gaps, whose sizes (and ages) were 186 [m.sup.2] (7 years), 266 [m.sup.2] (4 years), 313 [m.sup.2] (10 years), and 322 [m.sup.2] (7 years). These contained 1, 2, 1, and 1 C. schreberiana stems, respectively. Some stems had reached 10 m in height in the older gaps. Four C. schreberiana occurred in transition areas between the gap and the closed canopy forest, and one small stem was nearby under closed forest.
In another study I found C. schreberiana in some smaller gaps (N. Brokaw, unpubl. data). In the eight months prior to Hurricane Hugo, I located all gaps less than about 3 years old and [is greater than or equal to] 20 [m.sup.-2] in about 35 ha of tabonuco forest at El Verde, and I counted the C. schreberiana within them. There were 44 gaps in the area, ranging from 20 to 117 [m.sup.2] (N. Brokaw, unpubl. data). Eight contained C. schreberiana. A recently created gap of 107 [m.sup.2] contained 31 C. schreberiana, the tallest individual being 1.85 m. The other seven gaps contained 25 more C. schreberiana, from seedlings to a stem 9.5 m tall. I found a few saplings directly beneath closed canopy but near gaps or stream openings from which they received side light. Many of the gap-colonizing trees had attained higher-light environments by germinating on the fallen log or a tip-up mound.
Thus C. schreberiana does regenerate in background gaps, but in the Luquillo forests these gaps tend to be relatively fewer or smaller than in many other tropical forests and there- fore provide relatively little opportunity for regeneration of this species. In other neotropical forests gaps occur at rates of about 1 [ha.sup.-1] [yr.sup.-1] and average 100 [m.sup.2] per gap (Hartshorn, 1990). But at El Verde I estimated that gaps were occurring at a rate of about 0.5 [ha.sup.-1] [yr.sup.-1] and averaging 49.9 [m.sup.2] (s.e. = 6.63, n = 44) in the three years before Hurricane Hugo (N. Brokaw, unpubl. data). The largest gap was only 117 [m.sub.2]. In 13 ha of tabonuco forest in the Bisley Watersheds (Scatena & Lugo, 1995), gaps occurred at a rate of 0.8 [ha.sup.-1] [yr.sup.-1] during the two years before Hugo and averaged 76 [m.sup.2] (S.E. = 22, n = 19). The Bisley figures are not greatly lower than figures typical for tropical forests, but nearly half the gaps (9 of 19) and most of the gap area (936 of 1444 [m.sup.2]) occurred on unstable soils of riparian valleys, representing only 6.9% of the study area. On slopes and in upland valleys (76.6% of the study area), gap dynamics were similar to those at El Verde: 0.5 gaps [ha.sup.-1] [yr.sup.-1], averaging 56 [m.sup.2] (S.E. = 21, n = 6). On ridges (16.9% of the area), a small sample suggested that gaps were relatively frequent (0.9 [ha.sup.-1] [yr.sup.-1]) but small (55 [m.sup.2], S.E. = 45, n = 4).
Because background treefall gaps are generally few and small, regeneration in gaps is not sufficient to account for the densities of adult C. schreberiana in tabonuco forest. In my pre-hurricane survey of all gaps in 35 ha at El Verde, I found 1.6 seedling and sapling C. schreberiana per [ha.sup.-1] of forest. Yet, before Hurricane Hugo there were 8.5 C. schreberiana of [is greater than or equal to] 10 cm dbh per [ha.sup.-1] in the 16 ha Hurricane Recovery Plot (HRP) (Zimmerman et al., 1994) located within my 35 ha study area. To maintain 8.5 adult trees per [ha.sup.-1] with 1.6 juveniles per [ha.sup.-1] requires that adults be exceptionally long-lived and juveniles have exceptionally high survival, neither of which characterizes C. schreberiana (Silander, 1979).
Hurricanes have a tremendous impact on the Cecropia schreberiana population in the Luquillo Mountains. Hurricanes kill adults but promote seed germination and seedling growth. Forest damage due to individual hurricanes varies over the Luquillo forest landscape (Boose et al., 1994), but these storms are frequent enough to be an important key to understanding the species' population and size-class structures throughout the forest, and to have shaped the species' adaptations (Lugo & Scatena, 1996). In this century the Luquillo forests have been repeatedly damaged by hurricanes: moderately to severely in 1928 and 1931, severely in 1932, lightly in 1956, and moderately by Hugo in 1989 (Scatena & Larsen, 1991). Hurricanes of moderate intensity, such as Hugo, have passed over the Luquillo Mountains on average every 50-60 years (Scatena & Larsen, 1991). Hugo damaged the canopy in all forest types, producing extensive gaps suitable for C. schreberiana colonization. For instance, in hectare-sized study sites of tabonuco, colorado, and dwarf forest, Hugo increased the amount of forest area with vegetation [is less than or equal to] 2 m high from 0.4 to 26.1%, 2 to 27.3%, and 8.4 to 50.6%, respectively (Brokaw & Grear, 1991). (For more information on hurricane effects in Luquillo forests, see Walker et al., 1991, 1996a; Boose et al., 1994.)
Hurricanes kill a much larger proportion of Cecropia schreberiana adults than of most other species. As mentioned above, Hurricane Hugo killed 52.9% (72 of 136) of C. schreberiana of [is greater than or equal to] 10 cm dbh in the HRP at El Verde, whereas the mean mortality of all other common species in the study was only 8.4% (Zimmerman et al., 1994), and even after 171 weeks the hurricane-caused mortality of all trees of [is greater than or equal to] 5 cm dbh in 20 study plots at El Verde was only 13.3% (Walker, 1995). Elsewhere in the Luquillo Mountains, Hugo killed 87.5% (7 of 8) of C. schreberiana of [is greater than or equal to] 10 cm dbh in a 1 ha plot, compared to 24% mortality for all species in the plot (Dallmeier et al., 1992), and it caused much greater mortality to C. schreberiana than to other adventive species in a mahogany (Swietenia macrophylla King) plantation (Fu et al., 1996).
However, hurricanes also promote a flush of Cecropia schreberiana germination and seedling growth. In a tabonuco forest area, C. schreberiana was the third most abundant tree species in seedling (trees [is less than or equal to] 200 cm tall) plots studied after Hurricane Hugo (Guzman-Grajales & Walker, 1991). In another series of plots distributed among variously disturbed sites in the same forest, C. schreberiana seedlings (plants 10-100 cm tall) occurred at a mean density of 0.85 [m.sup.-2] nine months after the hurricane (Cammack, 1994). The peak number may have occurred earlier, but this was still much more than the roughly 0.00016 [m.sup.-2] C. schreberiana seedlings and saplings I had found in this forest before the storm (see above). About 40 months after the hurricane there were 10,635 (0.066 [m.sup.-2]) C. schreberiana of 1-10 cm dbh in the HRP (no pre-hurricane data for this size class) and a total of 565 of [is greater than or equal to]10 cm dbh, compared to 136 pre-hurricane stems of [is greater than or equal to] 10 cm dbh (Zimmerman et al., 1994; J. Thompson, unpubl. data; for more data on post-hurricane populations, see Fu et al., 1996).
Not surprisingly, high-light areas had high densities of Cecropia schreberiana after the hurricane (Fernandez & Fetcher, 1991), while presumed elevated soil nutrients from decomposing plant material following hurricanes (Sanford et al., 1991) would also favor its colonization. As mentioned, post-hurricane tip-up mounds and soil pits in tabonuco forest were especially dense with C. schreberiana (L. Walker, unpubl. data). There were a minimum of 1.42 C. schreberiana 10-100 cm tall per [m.sup.-2] in a sample of 27 soil pits and mounds (despite low soil fertility of pits) and none in the adjacent area of undisturbed soil. Saplings ([is greater than or equal to] 1 m tall) in these sites reached a mean density of 0.52 [m.sup.-2]. Cecropia schreberiana is also a common post-hurricane colonizer in colorado and palm forests (Weaver, 1989; Frangi & Lugo, 1991).
The positive impact of hurricanes on C. schreberiana establishment is evident from the fact that after Hurricane Hugo its abundance in natural forest and coffee plantations was inversely related to distance from the storm's path (Zimmerman et al., 1995). Similarly, C. schreberiana regeneration was more "apparent" in the heavily damaged Bisley area in the Luquillo Mountains than in the moderately damaged forest at El Verde after Hugo (Secrest et al., 1996).
Post-hurricane regeneration of C. schreberiana is substantial but occurs in a brief pulse (cf. Ferguson et al., 1995). Establishment ends within several months after a hurricane, as re-growth shades the understory (Fernandez & Fetcher, 1991) and prevents further germination from a seed bank depleted in any case (Nauman, 1991). After Hurricane Hugo the density of C. schreberiana seedlings peaked by nine months in the study by Cammack (1994) noted above, declining from 0.85 [m.sup.-2] to 0.03 [m.sup.-2] at 50 months post-hurricane, and mortality was high on tip-ups and in soil pits (L. Walker, unpubl. data). In the HRP the total population of C. schreberiana of [is greater than or equal to] 1 cm dbh declined from 11,200 at 40 months to 5392 at 84 months (J. Thompson, unpubl. data).
These data obtained after Hurricane Hugo carry the record of Cecropia schreberiana dynamics to seven years following a hurricane. To this account we can add records of C. schreberiana dynamics from 11 to 54 years after the severe Hurricane of 1932. In 1943 there were 39 C. schreberiana of [is greater than or equal to] 4 cm dbh in a 0.72 ha plot of hurricane-damaged tabonuco forest (Crow, 1980). By 1946 there were 50, making up 5.5% of the basal area of the plot. But by 1951 stem number had declined to 39, and in 1976 there were only six left. Cecropia schreberiana had the steepest survivorship curve among important species in the plot. Meanwhile, survivors had, of course, grown, so that the population, lacking recruitment, consisted of ever larger trees, and its share of plot basal area continued to increase for a short while even as stem number declined. In another tabonuco plot, of 0.4 ha, C. schreberiana decreased from 48 to 10 stems of [is greater than or equal to] 4 cm dbh between 1946 and 1976 (Weaver, 1994). Weaver (1983, 1986, 1989, 1990) reports similar steep declines in other tabonuco stands and in colorado and palm forest stands during the 1946-1981 period (and see Fu et al., 1996). As in Crow's (1980) study, C. schreberiana had the lowest survival during the period among important tree species in Weaver's (1983) sites. In other palm forest studies, C. schreberiana did not decline from 1946 to 1986 in one plot, for no given reason, but did decline in another (Lugo et al., 1995). Cecropia schreberiana is one of the first species to die out in developing palm forests, while the size of dying trees increases as the population ages (Lugo et al., 1995).
Post-hurricane dynamics of C. schreberiana are directly important to some animals. The coqui frog (Eleutherodactylus coqui) uses the large fallen leaves of C. schreberiana as nest sites, and in the five years after Hurricane Hugo this frog was especially abundant where C. schreberiana was abundant (Woolbright, 1996). Similarly, the lizard Anolis gundlachi was more abundant where it could use C. schreberiana saplings as understory perches (Reagan, 1991). However, the density of four snail species was not significantly correlated with "apparency" of C. schreberiana after Hugo (Secrest et al., 1996).
E. SUMMARY OF CECROPIA SCHREBERIANA POPULATION DYNAMICS
With this understanding of the biology of Cecropia schreberiana and its response to different kinds of disturbances I am now able to describe in a general way the long-term population dynamics of this species over the landscape of the Luquillo Mountains. Cecropia schreberiana colonizes road cuts, but these cover little area, and human disturbances otherwise contribute little to C. schreberiana populations. The species colonizes landslides, but landslides cover only a small fraction of the landscape. Background treefall gaps suitable for colonization are also uncommon. Together, these disturbances, plus streamsides, provide sites for a steady but low density of "background regeneration" by C. schreberiana in the Luquillo Mountains, and the resulting mature trees maintain the seed bank. However, these disturbances are not sufficient to account for the high abundance of C. schreberiana trees in Luquillo forests. Only hurricanes disturb the forest over enough area and often enough to promote the amount of regeneration that maintains C. schreberiana as one of the more important tree species. For example, various studies described above showed that the species was in serious decline in the El Verde area until Hurricane Hugo revived it, at least in tree form; it was no doubt still quite abundant in the soil seed bank.
Several workers have modeled the response of C. schreberiana to hurricanes. Doyle (1981) used information on its light requirement, growth, and longevity, to show that biomass of C. schreberiana in the Luquillo Mountains should peak at 15-30 years after a damaging hurricane, and that by 60 years only a few, sustained by what I call "background regeneration," would remain. His model accurately predicted C. schreberiana rank abundance in sample plots. O'Brien et al. (1992) showed how basal area of C. schreberiana, given the species' dependence on hurricane disturbance, would increase if global warming increases the frequency and intensity of hurricanes hitting Puerto Rico. Zimmerman et al. (1996) depicted curves showing various types of responses for ecosystem components within the first five post-hurricane years. A combination of their curve A (increase in first year, then decline), for C. schreberiana seedlings, and curve C (steep decrease then rise above pre-hurricane levels), for its larger trees, describes the C. schreberiana response.
I propose a graphic model of the long-term, landscape-level population dynamics of C. schreberiana that includes all life stages of the species and fully reflects its life history strategy as a pioneer with dormant seeds. Every tree species has at least one life-history phase in which it is unusually persistent, in order to survive periods unfavorable for reproduction or for recruitment to the reproductive class. Some species persist by extended life as seed-bearing adults, rare regeneration then being sufficient to maintain a population. Some species persist as suppressed seedlings and saplings, awaiting canopy opening. Others, such as C. schreberiana, persist by abundant seed production often coupled with facultative seed dormancy and germination cued by an indicator of canopy opening. A full model of a species' population dynamics will include its persistent phase. For C. schreberiana this means showing both the tree forms (seedling, sapling, adult) and the seed form and their dynamics relative to hurricanes (Fig. 3).
[Figure 3 ILLUSTRATION OMITTED]
In this model all the tree forms are combined in one curve and the soil seed population in another. The curve for trees is derived from the survivorship function given by Silander (1979) for a cohort that has colonized a disturbance (Fig. 2). Initial cohort size is set by seed availability in the soil. The curve for seeds in the soil is plotted from Silander's (1979) formula for age-specific output of viable seed per female tree (not taking into account senescence) and the estimated number of female trees in the cohort though time. For all tree forms combined, the abundance of C. schreberiana peaks within several months after a hurricane. Tree abundance then begins a long decline, at first steep then leveling off, as, lacking further disturbance and establishment, the modal tree life stage and size class shift from seedling to sapling to canopy tree. Finally, about 50% of canopy C. schreberiana are killed by the next hurricane (with negligible effect on model projections). Meanwhile, the population in seed form declined drastically at the time of the previous hurricane, because most seeds in the soil germinated, many reproductive trees were killed, and others were stripped of flowers and fruit. The seed population gradually builds back up as surviving trees again fruit, new recruits to the reproductive class begin fruiting, and, soon after the hurricane, forest regrowth inhibits germination. (These seeds are increasingly produced by fewer trees as the number of adults declines, but fecundity increases per tree with age, until senescence.) The extent of the decline in the C. schreberiana population in tree form and the increase in the seed form depends on the interval between hurricanes. At long intervals its abundance in both forms will level off, sustained only by background regeneration. Taking both tree and seed forms together, C. schreberiana is always common, unless the period between hurricanes is exceptionally long. The more severe the hurricane, the higher the subsequent peak of C. schreberiana abundance in tree form and the lower the nadir of its abundance in seed form (not shown in model). Hurricanes do not decrease or increase the population of C. schreberiana so much as they shift it back and forth between tree and seed forms.
V. Importance of Cecropia schreberiana
Because its population structure generally reflects the time elapsed since hurricane disturbance, the status of Cecropia schreberiana indicates the developmental status of Luquillo forests as a whole (Walker et al., 1996c). But, more importantly, as a rapidly and abundantly colonizing species, this most successional of canopy trees (Smith, 1970b) seems to play key roles in ecosystem function and in the development of forest structure and composition after disturbance.
Silander (1979) hypothesized that colonizing stands of C. schreberiana conserve nutrients in the recovering forest ecosystem by reducing runoff and leaching and by efficiently acquiring nutrients. Colonizing C. schreberiana did quickly concentrate nutrients after Hurricane Hugo. In heavily damaged tabonuco forests in the Bisley watersheds, the highest above-ground net primary productivity in the five years after Hugo occurred in the second year as a result of massive recruitment of C. schreberiana saplings to the [is greater than or equal to] 1.3 m tall size class (Scatena et al., 1996). This productivity was achieved as large amounts of nutrient-poor necromass in the watersheds were replaced by nutrient-rich tissue in fast-growing colonizers, C. schreberiana prominent among them, with particularly high concentrations of K and Mg in their foliage. Moreover, C. schreberiana's leaf litter decays relatively slowly (La Caro & Rudd, 1985), releasing these nutrients gradually; the species may similarly store nutrients in landslides and background treefall gaps. Thus it appears that C. schreberiana performs a key function at the plant-soil interface (Silver et al., 1996) at certain times and places in Luquillo forests, by having a disproportionately large role in capturing and storing nutrients from decomposing plants after disturbances.
Cecropia schreberiana may also facilitate successional development of forest structure and composition after disturbance in the Luquillo Mountains. Observations suggest that some disturbed areas are colonized either by C. schreberiana, leading to establishment of other tree species and succession to mature forest, or by a dense growth of ferns, herbaceous vines, or grasses (TFES, 1952) that excludes trees and hinders succession (Walker et al., 1996b). Which scenario occurs depends, presumably, on which colonizers establish first, either type being able to prevent later establishment of the other. There are no data as yet to evaluate this hypothesis. Cecropia schreberiana is generally uncommon in areas dense with ferns, vines, and grasses (Silander & Lugo, 1990), but in the understory of some forests severely disturbed by Hurricane Hugo, abundant, colonizing graminoids were replaced within a year by C. schreberiana and other pioneer trees (Walker, 1991), and C. schreberiana itself may inhibit tree colonization in the first years of landslide regeneration (Walker et al., 1996a).
If C. schreberiana performs key functions in community reorganization after disturbance, is it also a key species, not replaceable in these roles by any other tree species in this ecosystem? Or is there species redundancy for these functions--that is, lacking C. schreberiana, would other species play the same, efficient roles capturing nutrients and possibly facilitating succession? Cecropia schreberiana's often much greater abundance than other pioneer canopy species after disturbance suggests that it would have the most effect in these roles. For example, in the 16 ha HRP, at about 40 months post-hurricane there were 10,635 C. schreberiana of 1-10 cm dbh, compared to totals for other pioneer trees of 4101 Didymopanax morototoni (Aubl.) Decne. & Planch., 3780 Casearia arborea (L. C. Rich.) Urban, 1994 C. sylvestris Sw., and 973 Alchornea latifolia Sw. (J. Thompson, unpubl. data). As mentioned earlier, it was by far the most common woody species in a study of regeneration on landslides (Myster & Walker, 1997), and Walker (unpubl. data) pointed out the big potential influence of this species on tip-ups and in soil pits, where there was, on average, one C. schreberiana [is greater than or equal to] 1 m tall for every 2 [m.sup.2] of substrate about three years after Hurricane Hugo.
There are few tropical forest tree species for which one could write so detailed a description as I have for Cecropia schreberiana. Only for species of exceptional economic importance is there usually so much information, whereas Cecropia schreberiana has been studied for its ecological importance. Not only is it among the more abundant canopy tree species in the Luquillo Mountains and an indicator of forest development, but it also seems to perform, perhaps uniquely, key functions in ecosystem and community reorganization following disturbance. Nonetheless, there are important gaps in our understanding of the ecology and role of C. schreberiana. How does the density of C. schreberiana in the soil seed bank, so critical to its population dynamics, vary in space and time, with stand type and history, and with local abundance of mature C. schreberiana? Why is the species uncommon in abandoned pastures? Does C. schreberiana facilitate succession in the forest, and if so, how? And is it the only species in the Luquillo Mountains that could perform that role? Answers to these questions will tell us how this species maintains its high abundance and plays its important ecosystem roles. But, more broadly, understanding the biology of this one tree species, Cecropia schreberiana, which is so responsive to hurricanes, can help us understand much about the structure and function of a hurricane-dominated tropical forest.
I thank Frank McCormick, Ivette Perez Viera, Peter Weaver, Gisel Reyes, Shah Cammack, Win Everham, Randall Myster, Lawrence Walker, Tim Schowalter, Frank Wadsworth, Jill Thompson, and Jess Zimmerman for providing me with source material. Fred Scatena, Jess Zimmerman, Randall Myster, Peter Weaver, Ariel Lugo, Mitch Aide, and Ned Fetcher made helpful comments on the manuscript. Elizabeth Mallory drew Figures 2 and 3. Shawn Fraver helped with the Spanish abstract. Writing of this paper was supported by grant BSR-8811764 from the National Science Foundation to the Institute for Tropical Ecosystem Studies, University of Puerto Rico, and to the International Institute of Tropical Forestry, as part of the Long-Term Ecological Research Program in the Luquillo Experimental Forest. The U.S. Forest Service (Department of Agriculture) and the University of Puerto Rico provided additional support.
VIII. Literature Cited
Aide, T. M, J. K. Zimmerman, L. Herrera, M. Rosario & M. Serrano. 1995. Forest recovery in abandoned tropical pastures in Puerto Rico. Forest Ecol. Managem. 77: 77-86.
--,--, M. Rosario & H. Marcano. 1996. Forest recovery in abandoned cattle pastures along an elevational gradient in northeastern Puerto Rico. Biotropica 28: 537-548.
Alvarez-Buylla, E. R. & M. Martinez-Ramos. 1992. Demography and allometry of Cecropia obtusifolia, a neotropical pioneer tree--An evaluation of the climax-pioneer paradigm for tropical rain forests. J. Ecol. 80: 275-290.
Bates, C. Z. 1929. Efectos del huracan del 13 de Septiembre de 1928 en distintos arboles. Rev. Agric. Puerto Rico 23: 113-117.
Bell, C. R. 1970. Seed distribution and germination experiment. Pages D177-D182 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Bloomfield, J. 1993. Nutrient dynamics and the influence of substrate quality on the decomposition of leaves and fine roots of selected tree species in a lower montane tropical rain forest in Puerto Rico. Ph.D. thesis, Yale University, New Haven.
Boose, E. R., D. R. Foster & M. Fluet. 1994. Hurricane impacts to tropical and temperate forest landscapes. Ecol. Monogr. 64: 369-400.
Briscoe, C. B. & F. H. Wadsworth. 1970. Stand structure and yield in the tabonuco forest of Puerto Rico. Pages B79-B89 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Brokaw, N. V. L. 1985. Treefalls, regrowth, and community structure in tropical forests. Pages 53-69 in S. T. A. Pickett & P. S. White (eds.), The ecology of natural disturbance and patch dynamics. Academic Press, New York.
--. 1986. Seed dispersal, gap colonization, and the case of Cecropia insignis. Pages 323-331 in A. Estrada & T. H. Fleming (eds.), Frugivory and seed dispersal. W. Junk, The Hague.
-- & J. S. Grear. 1991. Forest structure before and after Hurricane Hugo at three elevations in the Luquillo Mountains, Puerto Rico. Biotropica 23: 386-392.
Brown, S., A. E. Lugo, S. Silander & L. Liegel. 1983. Research history and opportunities in the Luquillo Experimental Forest. U.S.D.A. Forest Service General Technical Report SO-44, Southern Forest Experiment Station, New Orleans, LA.
Byer, M. D. & P. L. Weaver. 1977. Early secondary succession in an elfin woodland in the Luquillo Mountains of Puerto Rico. Biotropica 9: 35-47.
Calderon, F. 1993. The role of mycorrhizae in the nutrient absorptive strategy of important landslide colonizers. M.S. thesis, University of Puerto Rico, Rio Piedras.
Cammack, S. E. 1994. Seedling recruitment and growth on hurricane-disturbed plots in a subtropical wet forest in Puerto Rico: The role of abiotic influences in the regeneration niche. M.S. thesis, University of Georgia, Athens.
Charles-Dominique, P. 1986. Inter-relations between frugivorous vertebrates and pioneer plants: Cecropia, birds, and bats in French Guyana. Pages 119-135 in A. Estrada & T. H. Fleming (eds.), Frugivory and seed dispersal. W. Junk, The Hague.
Crow, T. R. 1980. A rain forest chronicle: A 30-year record of change in structure and composition at El Verde, Puerto Rico. Biotropica 12: 42-55.
-- & D. F. Grigal. 1979. Numerical analysis of arborescent communities in the rain forest of the Luquillo Mountains, Puerto Rico. Vegetatio 40: 135-146.
-- & P. L. Weaver. 1977. Tree growth in a moist tropical forest of Puerto Rico. U.S.D.A., Forest Service Research Paper ITF-22.
Dallmeier, F., C. M. Taylor, J. C. Mayne, M. Kabel & R. Rice. 1992. Case study of SI/MAB biological diversity plot research methodology: Effects of Hurricane Hugo on the Bisley Biodiversity Plot, Luquillo Biosphere Reserve, Puerto Rico. Pages 47-72 in R. Dallmeier (ed.), Long-term monitoring of biological diversity in tropical forest areas: Methods for establishment and inventory of permanent plots. MAB Digest 11, UNESCO, Paris.
Denslow, J. S. 1980. Gap partitioning among tropical rainforest trees. Biotropica 12: 47-55.
Devoe, N. N. 1989. Differential seeding and regeneration in openings and beneath closed canopy in subtropical wet forest. Ph.D. thesis, Yale University, New Haven.
Doyle, T. W. 1981. The role of disturbance in the gap dynamics of a montane rain forest: An application of a tropical forest succession model. Pages 56-73 in D. C. West et al. (eds.), Forest succession: Concepts and application. Springer-Verlag, New York.
Duke, J. A. 1970a. Keys for the identification of seedlings of some prominent woody species in eight forest types in Puerto Rico. Pages B239-B274 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
--. 1970b. Postirradiation woody seedlings. Pages D189-D192 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Edmisten, J. 1970. Survey of mycorrhiza and nodules in the El Verde forest. Pages F 15-F20 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Estrada Pinto, A. 1970. Phenological studies of trees at El Verde. Pages D237-D269 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Everham, E. M., III, R. W. Myster & E. VanDeGenachte. 1996. Effects of light, moisture, temperature, and litter on the regeneration of five tree species in the tropical montane wet forest of Puerto Rico. Amer. J. Bot. 83: 1063-1068.
Ewel, J. J. & J. L. Whitmore. 1973. The ecological life zones of Puerto Rico and the U.S. Virgin Islands. U.S.D.A. Forest Service Research Paper ITF-18. Institute of Tropical Forestry, Rio Piedras, Puerto Rico.
Ferguson, B. G., D. H. Boucher, M. Pizzi & C. Rivera. 1995. Recruitment and decay of a pulse of Cecropia in Nicaraguan rain forest damaged by Hurricane Joan: Relation to mutualism with Azteca ants. Biotropica 27: 455-460.
Fernandez, D. S. & N. Fetcher. 1991. Changes in light availability following Hurricane Hugo in a subtropical montane forest in Puerto Rico. Biotropica 23: 393-399.
Fetcher, N., B. L. Haines, R. A. Cordero, D. J. Lodge, L. R. Walker, D. S. Fernandez & W. T. Lawrence. 1996. Responses of tropical plants to nutrients and light on a landslide in Puerto Rico. J. Ecol. 84: 331-341.
Fleming, T. H. & E. R. Heithaus. 1981. Frugivorous bats, seed shadows, and the structure of tropical forests. Biotropica 13: 45-53.
Francis, J. K. & A. J. R. Gillespie. 1993. Relating gust speed to tree damage in Hurricane Hugo, 1989. J. Arboricult. 19: 368-373.
Frangi, J. L. & A. E. Lugo. 1991. Hurricane damage to a flood plain forest in the Luquillo Mountains of Puerto Rico. Biotropica 23: 324-335.
Fu, S., C. Rodriguez Pedraza & A. E. Lugo. 1996. A twelve year comparison of stand changes in a mahogany plantation and a paired natural forest of similar age. Biotropica 26:515-524.
Garcia-Martino, A. R., G. S. Warner, F. N. Scatena & D. L. Civico. 1996. Rainfall, runoff, and elevation in the Luquillo Mountains of Puerto Rico. Caribbean J. Sci. 32: 413-424.
Garcia-Montiel, D. & F. N. Scatena. 1994. The effect of human activity on the structure and composition of a tropical forest in Puerto Rico. Forest Ecol. Managem. 63: 57-78.
Glogiewicz, J. In press. Cecropia peltata L. Useful trees of tropical North America. U.S.D.A. Forest Service, Washington, DC.
Guariguata, M. R. 1990. Landslide disturbance and forest regeneration in the upper Luquillo Mountains of Puerto Rico. J. Ecol. 78: 814-832.
--& M. C. Larsen. 1990. Preliminary map showing locations of landslides in El Yunque Quadrangle, Puerto Rico. U.S. Geological Survey Open-File Report 89-257.
Guzman-Grajales, S. M. & L. R. Walker. 1991. Differential seedling responses to litter after Hurricane Hugo in the Luquillo Experimental Forest, Puerto Rico. Biotropica 23: 407-413.
Hartshorn, G. S. 1990. An overview of neotropical forest dynamics. Pages 585-599 in A. H. Gentry (ed.), Four neotropical rainforests. Yale University Press, New Haven.
Henderson, A., G. Galeano & R. Bernal. 1995. Field guide to the palms of the Americas. Princeton University Press, Princeton.
Holdridge, L. R. 1970. A system for representing structure in tropical forest associations. Pages B147-B150 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Holthuijzen, A. A. M. & J. H. A. Boerboom. 1982. The Cecropia seed bank in the Surinam lowland rain forest. Biotropica 14: 62-68.
Howard, R. A. 1988. Flora of the Lesser Antilles: Leeward and Windward Islands. Vol. 4: Dicotyledoneae--Part 1. Arnold Arboretum, Jamaica Plain, MA.
ISTF (International Society of Tropical Foresters). 1997. Cecropia name changed. ISTF News 18: 4.
ITF (Institute of Tropical Forestry). 1959. 1958 annual report: Tropical forest research center. Caribbean Forest. 20: 1-10.
Janzen, D. H. 1973. Dissolution of mutualism between Cecropia and its Azteca ants. Biotropica 5: 15-28.
La Caro, F. & R. L. Rudd. 1985. Leaf litter disappearance rates in Puerto Rican montane rain forest. Biotropica 17: 269-276.
Larsen, M. C. & A. J. Torres-Sanchez. 1992. Landslides triggered by Hurricane Hugo in eastern Puerto Rico, September 1989. Caribbean J. Sci. 28(3/4): 13-120.
Lebron, M. L. 1977. An autecological study of Palicourea riparia Benth. (Rubiaceae), an ecologically important species in the recovery of a disturbed tropical rain forest in Puerto Rico. Ph.D. dissertation, University of North Carolina, Chapel Hill.
Leck, C. F. 1972. Observations of birds at Cecropia trees in Puerto Rico. Wilson Bull. 84: 498-500.
Liogier, H. A. 1985. Descriptive flora of Puerto Rico and adjacent islands. Vol. 1. Editorial de la Universidad de Puerto Rico, Rio Piedras.
Little, E. L., Jr. & F. H. Wadsworth. 1964. Common trees of Puerto Rico and the Virgin Islands. Agriculture Handbook No. 249, U.S.D.A. Forest Service, Washington, DC.
Lodge, D. J. 1996. Microorganisms. Pages 53-108 in D. P. Reagan & R. B. Waide (eds.), The food web of a tropical rain forest. University of Chicago Press, Chicago.
Lugo, A. 1970. Photosynthetic studies on four species of rain forest seedlings. Pages I81-I102 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
-- & F. N. Scatena. 1995. Ecosystem-level properties of the Luquillo Experimental Forest with emphasis on the tabonuco forest. Pages 59-108 in A. E. Lugo & C. Lowe (eds.), Tropical forests: Management and ecology. Springer-Verlag, Berlin.
-- & --. 1996. Background and catastrophic tree mortality in tropical moist, wet, and rain forests. Biotropica 28: 585-589.
--, A Bokkestijn & F. N. Scatena. 1995. Structure, succession, and soil chemistry of palm forests in the Luquillo Experimental Forest. Pages 142-177 in A. E. Lugo & C. Lowe (eds.), Tropical forests: Management and ecology. Springer- Verlag, Berlin.
Mabberly, D. J. 1987. The plant book: A portable dictionary of the higher plants. Cambridge University Press, Cambridge.
Marrero, J. 1954. Regeneration: Seed studies. Cecropia peltata. U.S.D.A. Forest Service, Institute of Tropical Forestry report. Rio Piedras, Puerto Rico.
McCormick, J. F. 1970. Direct and indirect effects of gamma radiation on seedling diversity and abundance in a tropical forest. Pages D-201-D-211 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
--. 1995. A review of the population dynamics of selected tree species in the Luquillo Experimental Forest, Puerto Rico. Pages 224-257 in A. E. Lugo & C. Lowe (eds.), Tropical forests: Management and ecology. Springer-Verlag, Berlin.
McKey, D. 1988. Cecropia peltata, an introduced neotropical pioneer tree, is replacing Musanga cecropioides in southwestern Cameroon. Biotropica 20: 262-264.
Murphy, P. G. 1970. Tree growth at El Verde and the effects of ionizing radiation. Pages D141-D171 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Myster, R.W. 1994. Landslide insects show small differences between an island (Puerto Rico) and the mainland (Costa Rica). Acta Cient. 8:105-113.
--. 1997. Seed predation, disease and germination on landslides in neotropical lower montane wet forest. J. Veg. Sci. 8: 55-64.
-- & L. R. Walker. 1997. Plant successional pathways on Puerto Rican landslides. J. Trop. Ecol. 13: 165-173.
Nauman, M. 1991. Seed bank in soils of Bisley watersheds. Inst. Trop. Forest. Ann. Lett. 1990-1991: 17.
O'Brien, S. T, B. P. Hayden, & H. H. Shugart. 1992. Global climatic change, hurricanes, and a tropical forest. Climat. Change 22: 175-190.
Odum, H. T. 1970a. The AEC rain forest program. Pages C3-C22 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
--.1970b. Summary: An emerging view of the ecological system at El Verde. Pages 1191-I289 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
-- & G. Cintron. 1970. Forest chlorophyll and radiation. Pages I21-I33 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
-- & R. F. Pigeon (eds.). 1970. A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
--, A. Lugo, G. Cintron, & C. F. Jordan. 1970a. Metabolism and evapotranspiration of some rain forest plants and soil. Pages I103-I164 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
--, P. Murphy, G. Drewry, F. McCormick, C. Sehinan, E. Morales & J. A. McIntyre. 1970b. Effects of gamma radiation on the forest at El Verde. Pages D3-D75 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Ogle, K. J. 1970. Pollen analysis of selected sphagnum bog sites in Puerto Rico. Pages B 135-B 145 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Olson, S. L. & K. E. Blum. 1968. Avian dispersal of plants in Panama. Ecology 49: 565-566.
Ovington, J. D. & J. S. Olson. 1970. Biomass and chemical content of El Verde lower montane rain forest plants. Pages H53-H77 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Perez Viera, I. E. 1986. Tree regeneration in two tropical rain forests. M.S. thesis, University of Puerto Rico, Rio Piedras.
Reagan, D. P. 1991. The response of Anolis lizards to hurricane-induced habitat changes in a Puerto Rican rain forest. Biotropica 23: 468-474.
--. 1996. Anoline lizards. Pages 321-345 in D. P. Reagan & R. B. Waide (eds.), The food web of a tropical rain forest. University of Chicago Press, Chicago.
-- & R. B. Waide (eds.). 1996. The food web of a tropical rain forest. University of Chicago Press, Chicago.
Rickson, F. R. 1977. Progressive loss of ant-related traits of Cecropia peltata on selected Caribbean islands. Amer. J. Bot. 64: 585-592.
Sanford, R. L., Jr., W. J. Parton, D. S. Ojima & D. J. Lodge. 1991. Hurricane effects on soil organic matter dynamics and forest production in the Luquillo Experimental Forest, Puerto Rico: Results of simulation modeling. Biotropica 23: 364-372.
Scatena, F. N. 1989. An introduction to the physiology and history of the Bisley Experimental Watersheds in the Luquillo Mountains of Puerto Rico. General Technical Report SO-72. U.S.D.A. Forest Service, Southern Forest Experiment Station, New Orleans, LA.
-- & M. C. Larsen. 1991. Physical aspects of Hurricane Hugo in Puerto Rico. Biotropica 23: 317-323.
-- & A. E. Lugo. 1995. Geomorphology, disturbance, and the soil and vegetation of two subtropical wet steepland watersheds of Puerto Rico. Geomorphology 13: 199-213.
--, W. Silver, T. Siccama, A. Johnson & M. J. Sanchez. 1993. Biomass and nutrient content of the Bisley Experimental Watersheds, Luquillo Experimental Forest, Puerto Rico, before and after Hurricane Hugo. Biotropica 25:15-27.
--, S. Moya, C. Estrada & J. D. Chinea. 1996. The first five years in the reorganization of aboveground biomass and nutrient use following Hurricane Hugo in the Bisley Experimental Watersheds, Luquillo Experimental Forest, Puerto Rico. Biotropica 28: 424-440.
Schowalter, T. D. 1994. Invertebrate community structure and herbivory in a tropical rain forest canopy in Puerto Rico following Hurricane Hugo. Biotropica 26:312-319.
--. 1995. Canopy invertebrate community response to disturbance and consequences of herbivory in temperate and tropical forests. Selbyana 16: 41-48.
--. 1997. Canopy invertebrate community response to hurricane disturbance. Luquillo LTER Annual Report. Institute for Tropical Ecosystem Studies, University of Puerto Rico, San Juan.
Scogin, R. 1982. Dietary observations on the red, fig-eating bat (Stenoderma rufum) in Puerto Rico. Aliso 10: 259-261.
Secrest, M. F., M. R. Willig & L. L. Peppers. 1996. The legacy of disturbance on habitat associations of terrestrial snails in the Luquillo Experimental Forest, Puerto Rico. Biotropica 28:502-514.
Silander, S. R. 1979. A study of the ecological life history of Cecropia peltata L., an early secondary successional species in the rain forest of Puerto Rico. M.S. thesis, University of Tennessee, Knoxville.
-- & A. E. Lugo. 1990. Cecropia peltata L.: Yagrumo hembra, trumpet tree. Pages 244-249 in Silvics of North America. Vol. 2: Hardwoods. Agricultural Handbook 654, U.S. Department of Agriculture, Washington, DC.
Silver, W. L., S. R. Brown & A. E. Lugo. 1996. Effects of changes in biodiversity on ecosystem function in tropical forests. Conserv. Biol. 10: 17-24.
Smith, R. F. 1970a. Preliminary illustrated leaf key to the woody plants of the Luquillo Mountains. Pages B275-B290 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
--. 1970b. The vegetation structure of a Puerto Rican rain forest before and afar short-term gamma radiation. Pages D103-D140 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Snyder, N. F. R., J. W. Wiley & C. B. Kepler. 1988. The parrots of Luquillo: Natural history and conservation of the Puerto Rican parrot. Western Foundation of Vertebrate Zoology, Los Angeles.
Soriano-Ressy, M., A. P. Desmarais & J. W. Perez. 1970. A comparison of environments of rain forests in Dominion, British West Indies, and Puerto Rico. Pages B329-B346 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Taylor, C. M., S. Silander, R. B. Waide & W. J. Pfeiffer. 1995. Recovery of a tropical forest after gamma irradiation: A 23-year chronicle. Pages 258-285 in A. E. Lugo & C. Lowe (eds.), Tropical forests: Management and ecology. Springer- Verlag, Berlin.
TFES (Tropical Forest Experiment Station, U.S. Forest Service). 1952. Twelfth Annual Report. Caribbean Forest. 13: 1-21.
Vasquez-Yanes, C. & H. Smith. 1982. Phytochrome control of seed germination in the tropical rain forest pioneer trees Cecropia obtusifolia and Piper auritum and its ecological significance. New Phytol. 92: 477-485.
Wadsworth, F. H. 1950. Notes on the climax forests of Puerto Rico and their destruction and conservation prior to 1900. Caribbean Forest. 11: 38-47.
--. 1951. Forest management in the Luquillo Mountains. I. The setting. Caribbean Forest. 12: 93-114.
--. 1957. Seventeenth annual report of the tropical forest research center. Caribbean Forest. 18:1-11.
--. 1958. The status of forestry and forestry research in Puerto Rico and the Virgin Islands: The eighteenth annual report of the Tropical Forest Research Center. Caribbean Forest. 19: 1-24.
--. 1970. Review of past research in the Luquillo Mountains of Puerto Rico. Pages B33-B46 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Wadsworth, R. K. 1970. Point-quarter sampling of forest type-site relations at El Verde. Pages B97-B 104 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
Walker, L. R. 1991. Tree damage and recovery from Hurricane Hugo in Luquillo Experimental Forest, Puerto Rico. Biotropica 23: 379-385.
--. 1995. Timing of post-hurricane tree mortality in Puerto Rico. J. Trop. Ecol. 11:315-320.
--. 1994. Effects of fern thickets on woodland development on landslides in Puerto Rico. J. Veg. Sci. 5: 525-532.
-- & L. E. Neris. 1993. Posthurricane seed rain dynamics in Puerto Rico. Biotropica 25: 408-418.
--, N. V. L. Brokaw, D. J. Lodge & R. B. Waide (eds.). 1991. Special issue: Ecosystem, plant, and animal responses to hurricanes in the Caribbean. Biotropica 23:313-521.
--, W. L. Silver, M. R. Willig & J. K. Zimmerman (eds.). 1996a. Special issue: Long-term responses of Caribbean ecosystems to disturbance. Biotropica 28:413-613.
--, D. J. Zarin, N. Fetcher, R. W. Myster & A. H. Johnson. 1996b. Ecosystem development and plant succession on landslides in the Caribbean. Biotropica 28: 566-576.
--, J. K. Zimmerman, D. J. Lodge & S. Guzman-Grajales. 1996c. An altitudinal comparison of growth and species composition in hurricane-damaged forests in Puerto Rico. J. Ecol. 84: 877-889.
Weaver, P. L. 1979. Tree growth in several tropical forests of Puerto Rico. Research Paper SO-152. U.S.D.A. Forest Service, Southern Forest Experiment Station, New Orleans.
--. 1983. Tree growth and stand changes in the subtropical life zones of the Luquillo Mountains of Puerto Rico. Research Paper SO-190. U.S.D.A. Forest Service, Southern Forest Experiment Station, New Orleans.
--. 1986. Hurricane damage and recovery in the montane forests of the Luquillo Mountains of Puerto Rico. Caribbean J. Sci. 22: 53-70.
--. 1989. Forest changes after hurricanes in Puerto Rico's Luquillo Mountains. Interciencia 14: 181-192.
--. 1990. Succession in the elfin woodland of the Luquillo Mountains of Puerto Rico. Biotropica 22: 83-89.
--. 1991. Environmental gradients affect forest composition in the Luquillo Mountains of Puerto Rico. Interciencia 16:142-151.
--. 1994. Bano de Oro Natural Area Luquillo Mountains, Puerto Rico. General Technical Report SO-111. U.S.D.A. Forest Service, Southern Forest Experiment Station, New Orleans.
--. 1995. The colorado and dwarf forests of Puerto Rico's Luquillo Mountains. Pages 109-141 in A. E. Lugo & C. Lowe (eds.), Tropical forests: Management and ecology. Springer-Verlag, Berlin.
-- & P. G. Murphy. 1990. Forest structure and productivity in Puerto Rico's Luquillo Mountains. Biotropica 22: 69-82.
--, E. Medina, D. Pool, K. Dugger, J. Gonzales-Liboy & E. Cuevas. 1986. Ecological observations in the dwarf cloud forest of the Luquillo Mountains in Puerto Rico. Biotropica 18: 79-85.
Weinbren, M. P., B. M. Weinbren, W. B. Jackson & J. B. Villella. 1970. Studies on the roof rat (Rattus rattus) in the El Verde forest. Pages E 169-E 181 in H. T. Odum & R. F. Pigeon (eds.), A tropical rain forest: A study of irradiation and ecology at El Verde, Puerto Rico. National Technical Information Service, Springfield, VA.
White, H. H. 1963. Variation of stand structure correlated with altitude in the Luquillo Mountains. Caribbean J. Forest. 1: 46-52.
Whitmore, T. C. 1990. An introduction to tropical rainforests. Clarendon Press, Oxford.
Willig, M. R. & M. R. Gannon. 1996. Mammals. Pages 399-431 in D. P. Reagan & R. B. Waide (eds.), The food web of a tropical rain forest. University of Chicago Press, Chicago.
Woolbright, L. 1996. Disturbance influences long-term population patterns in the Puerto Rican frog, Eleutherodactylus coqui (Anura: Leptodactylidae). Biotropica 28: 493-501.
Zimmerman, J. K., E. M. Everham III, R. B. Waide, D. J. Lodge, C. M. Taylor & N. V. L. Brokaw. 1994. Responses of tree species to hurricane winds in subtropical wet forest in Puerto Rico: Implications for tropical tree life histories. J. Ecol. 82:911-922.
--, T. M. Aide, M. Rosario, M. Serrano & L. Herrera. 1995. Effects of land management and a recent hurricane on forest structure and composition in the Luquillo Experimental Forest, Puerto Rico. Forest Ecol. Managem. 77: 65-76.
--, M. R. Willig, L. R. Walker & W. L. Silver. 1996. Introduction: Disturbance and Caribbean ecosystems. Biotropica 28: 414423.
Zou, X., C. P. Zucca, R. B. Waide & W. H. McDowell. 1995. Long-term influence of deforestation on tree species composition and litter dynamics of a tropical rain forest in Puerto Rico. Forest Ecol. Managem. 78: 147-157.
|Printer friendly Cite/link Email Feedback|
|Author:||Brokaw, Nicholas V.L.|
|Publication:||The Botanical Review|
|Date:||Apr 1, 1998|
|Previous Article:||Phylogenetics and character evolution in the grass family (Poaceae): simultaneous analysis of morphological and chloroplast DNA restriction site...|
|Next Article:||CAM photosynthesis in submerged aquatic plants.|